Verizon Takes Ericsson AI RAN Controller Nationwide This Year

Verizon will deploy Ericsson's Intelligent Controller (EIC) across its entire US network this year, handing the operator a nationwide platform for AI-native rApps that automate the radio access network. The announcement comes with field numbers attached, which is rarer in network AI than it should be.

In trials ahead of the rollout, fully autonomous closed-loop rApps ran round the clock and delivered measurable gains. Ericsson's Advanced Antenna System Cell Shaper raised downlink data radio bearer throughput by up to 8.9% and uplink UE throughput by up to 8%. Those figures come from the companies themselves, so treat them as vendor-and-operator reporting rather than independent measurement. Still, they're numbers, and network AI stories don't always offer them.

What does the controller actually do?

Matt Butz, Verizon's VP of network planning, said the work improves performance across both mobility and fixed wireless access. Building its own rApps on the open platform, he added, will let Verizon automate more of the network as it prepares for 6G. That's a planning statement, not a launch date, but it signals where the operator thinks autonomy fits its roadmap.

Verizon has also implemented a connected-mode handover method from LTE to standalone 5G for FWA devices. The practical win is modest and real: no more manual intervention when a handover fails. Anyone who has watched a fixed wireless customer complain about drops will appreciate that.

The operator plans to test more rApps, both its own and Ericsson's. The companies say this will accelerate progress toward Level 4 autonomy. On TM Forum's 0-to-5 scale, Level 4 means a highly autonomous network where the system makes decisions across a complex, cross-domain environment through predictive or active closed-loop management. In plain terms, the network fixes itself before anyone notices.

How close is Level 4, really?

Joe Constantine, CTO of Ericsson Americas, is careful here. He told Mobile World Live that Ericsson already has Level 4 products deployed with customers in many places, and vertical use cases exist at that level. An entire network is another matter.

"Vertical use case, yes, horizontal across the whole network, yet to be verified," he said. It's a refreshingly honest assessment from a vendor whose business depends on selling the vision.

Constantine argued full autonomy becomes necessary as networks start serving robots, drones and autonomous vehicles. A drone inspecting a power line must show it finished its mission, he noted, and the network needs to know it. "If you want to be intentional, you need autonomy," he said.

He also pointed to the openness of the platform. The rApps will work with anyone's software, he said, from Nokia to "two smart kids at university." That claim is worth watching as the ecosystem matures; open interfaces are easy to promise and harder to police.

Where the money and the data sit

So far, Constantine said, rApps have mostly been used for performance gains and energy savings. Power costs remain a genuine concern for operators, which explains why energy rApps keep topping deployment lists. A percentage of throughput is nice; a percentage off the electricity bill gets finance's attention.

More broadly, he predicted the winners in network AI will be the companies with the best access to trusted data. That's a quieter claim than most AI rhetoric, and probably the most consequential one in the story. Scale operators with clean data pipelines hold the advantage.

The EIC sits within Ericsson's Intelligent Automation Platform, which supports more than 100 rApps from over 100 members. That's the ecosystem Verizon is plugging into nationally this year, with more rApp testing to follow as the operator works toward higher autonomy levels ahead of 6G.